DESCRIPTION: (Applicant's Description) Prostate cancer will be diagnosed in more than 350,000 American men in 1996. More than 42,000 American men will die from prostate cancer in 1996. Although hormone therapy is effective initially in approximately 75% of the men with metastatic prostate cancer, this therapy is short-lived with virtually all men eventually developing progressive disease and dying of androgen-independent cancer. No effective systemic therapy exists beyond hormone therapy. Clearly new treatments are needed for androgen independent prostate cancer. Monoterpenes have a wide range of anticancer activity in rodent and in vitro models. These activities span prevention and treatment of multiple cancer types. The most extensively studied and impressive results have been in mammary carcinoma models where dietary monoterpene (perillyl alcohol) causes the complete regression of the majority of carcinomas in the same rodent models which led to the development of tamoxifen. The investigators have shown perillyl alcohol to be inhibitory in vitro to three human prostate cancer cell lines. In vivo, it effectively slows the growth of androgen independent human prostate cancer xenografts in mice. In addition, monoterpenes show a diverse array of metabolic, cellular and molecular activities both in vitro and in vivo, including inhibition of cellular proliferation, induction of differentiation and apoptoses and differential gene regulation of growth regulatory proteins. Finally, preliminary Phase I testing suggests that perillyl alcohol is well tolerated in cancer patients. The primary goal of this project is to assess the clinical effectiveness of daily perillyl alcohol in androgen independent prostate cancer patients enrolled in a Phase II study. Perillyl alcohol will be given in 3 divided doses daily. One additional pilot study will be performed in conjuncture with the Phase II study. Since a potential major mechanism for the antitumor activity of monoterpenes is the induction of the mannose-6-phosphate (M6P)/insulin-like growth factor (IGFII) receptor and transforming growth factor beta (TGFbeta) and since prior studies show TGFbeta inhibition of androgen independent prostate cancer cells, serum TGFbeta will be examined as a potential intermediate endpoint for perillyl alcohol activity.